I think xoggyux did a pretty good job in this post.
All the resources necessary to put the pieces together are, ironically, linked in the Macrumors article. The law
doesn't say what's reported in the NYT article, which says "the law states that
a variable will revert to a mean over a large sample of results".
This is
not true!
The law states roughly that
the average value in an increasingly large sample will, over time, tend to approach the true average. There's a crucial difference between making a claim about the behaviour of the
value of a variable (stock price, in this case), and the behaviour of
averages of a sample of such values. This is the content of the second sentence in the
linked Wikipedia article.
In this case, a correct application of the theorem would say something like "if you keep taking the average of the stock prices over an extended period of time, the longer you keep calculating average stock prices, the closer
that average will get to the true stock price [whatever that is!]."
I can't see how that supports the claims in the article.